EEVblog Electronics Community Forum
Products => Thermal Imaging => Topic started by: Ultrapurple on May 31, 2020, 02:54:04 pm
-
Those with a long memory will recall that some time ago I acquired a big, almost antique, Agema reflector lens and I hadn't had a great deal of luck getting useful images from it. I have posted progress (such as it has been) before.
Now, however, I am delighted to say I am getting proper images from it!
But I shouldn't be.
[attach=1]
(Click images to embiggen).
The lens is clearly marked, front and rear, "LW" - ie long wave, 7-14µm or thereabouts. But I recently acquired an Agema Thermovision 870 (a SPRITE-based semi-cooled MWIR camera) in fairly good, working condition. It came with a 20° MWIR lens. I was surprised to discover that a 7.5° LWIR lens I had lying around not only mounted, but also gave very good pictures. I wasn't expecting to see anything, of course, because if something is labelled "MWIR" its coatings normally won't pass LWIR (and vice-versa). Interestingly, the images from the 7.5° LWIR lens were pretty much as 'bright' as those from the 20° MWIR lens, leading me to think that they probably have broad-band coatings. (If the coating was 'leaking' some MWIR then I'd expect it to be much 'darker' than a true MWIR lens).
I realised that my big lens not only had the same kind of mount, but it also had a perfectly-placed support for the camera body as well: clearly, they would fit together. So I tried it.
[attach=2]
Yup - it all fits!
The minimum focus distance for the lens is stated as 20m (~60ft) so I pointed it at a neighbour's chimneys roughly 40m away.
[attach=3]
What do you know? I could get a nice, fairly clearly focused image!
[attach=4]
The '10' at the top of the image indicates a 10°C span.
I took a few more images of the same subject:
[attach=5]
And even managed to spot a bird on the chimney.
[attach=6]
We need to remember that these images were made by photographing the readout unit's CRT screen as it balanced precariously on a shelf in my shed:
[attach=7]
In order to reduce the issues caused by the 4:1 interlace I have been using long exposures - 3 to 4 seconds. This means arranging very dark surroundings and running the visible light camera at f/32 to f/36 (!) to stop the sensor peaking out. I have found that anything less than a second or so leads to significant emphasis on one or other of the fields (leading to one very visible set of lines that stand out from all the others). Some of the images in this post show the effect - it's much worse with a faster shutter.
I don't have anything that can produce a pure-electronic image (yet). The camera is a scanning-mirror design and its video standard is odd (6.25fps, 4:1 interlace, 25Hz fields) so I will have to construct something specially when I have time. (I have good ideas on how to do this). Meanwhile, here are some images I made a few days ago with the 20° MWIR lens looking at the readout unit and system PSU - note the way one raster is more visible, as discussed earlier:
[attach=8]
I think this is doing reasonably well for an early-1980s camera. Its resolution is open to debate but I believe it's supposed to be 280 lines (4 fields of 70) and, looking at the images, about 250 pixels per line - so call it 280V x 250H. There aren't any 'true' pixels, as it's a purely analogue system.
I shall continue experimenting, but it's really nice to know that I'm finally able to see something through the lens - and it does have ENORMOUS magnification!
-
Good to see it going.
I doubt any of the coatings & responses (detector or lens) are such a hard cut off between 5µm and 7µm. It also helps with the big lens having at least two reflective surfaces - is there any transmissive optic in there ?
Even so, if say the LW coating were say leaking 10% in the MW but LW energy content in that band is 10x the MW for ambient scenes, would that not give equal brightness as well ?
Bill
-
A mirror does not need a special coating for the IR. A surface coating is only for corrosion protection. Different surface coatings may make a difference at the blue and UV end.
Chances are high there is no IR active coating at all on the mirrors, just some thin "natural" aluminum oxide. So the mirrors could be good all the way from the visible to the microwave range.
-
A mirror does not need a special coating for the IR. A surface coating is only for corrosion protection. Different surface coatings may make a difference at the blue and UV end.
Chances are high there is no IR active coating at all on the mirrors, just some thin "natural" aluminum oxide. So the mirrors could be good all the way from the visible to the microwave range.
Yes, you're right, however this lens has a big lump of (coated) germanium just inside the lens mount. If I took that out I suspect the telescope would work fine from visible light downwards (although the coatings are a bit dirty with age and I haven't been able to clean them, preferring to leave them as they are rather than risk damage).
-
Today ..... roof tops......... tomorrow ....... the Moon ;D :-+
Fraser
-
I have given the moon some serious consideration.
The moon subtends about 0.5° visual angle, or one-fifth the lens' field of view. That would mean it would be the equivalent of about 50 'pixels' wide. If all went well I might be able to stretch this a bit with stacking etc but realistically it's unlikely to look much better than the old LWIR image I made (on the right), shown here approximately 50 pixels wide:
(https://live.staticflickr.com/8642/16495320867_72284bc256_o_d.jpg)
So one might be able to tell it was the moon, just...
I'm more interested in getting the lens to work with one or more of the higher resolution LWIR cameras in my arsenal. Now I know I can set the lens to good focus on an object, I can fix it looking at the target and then play with subsidiary optics and sensors in the knowledge that two variables (focus and 'a target') are fixed, meaning I can concentrate on adjusting the other 74 or so...
But now I know I can definitely set the lens up focused at point X, I have a fixed point from which to work when using the other cameras.
-
Hmm, What you need is the correct rear lens assembly from an AGEMA thermal camera designed for that reflector lens. Oh, I just happen to have one from my disassembled THV470 ;D
Sadly the AR coating has peeled due to age or moisture exposure. I am not sure how much the damage effects an image through the assembly though. I do have the original 20 degree AGEMA lens that was used with the mount so may be able to test it with LWIR and MWIR cameras.
But is it MWIR coated or broadband I wonder ? I will have to test it.
Fraser
-
Ok LWIR test completed.
The lens is a 12 degree model and not 20 degree. When mounted on the rear lens assembly it provided a good, but inverted, image of my coffee cup when placed in front of my FLIR E40 (E60+). The coffee cup had a side temperature of 42C but measured as 26C through the AGEMA lens. That is a significant transmission loss so these parts may well be coated for MWIR.
Fraser
-
I will dismantle the AGEMA rear lens assembly so we can see what elements are used in it. It appears to contain 2 elements but whether positive or negative is not known yet. The 12 degree removable lens that I have contains a single lens element. The output of the complete lens assembly may be accessed using a thermal camera that has manual focus, like the Exx series.
Fraser
-
I have partially dismantled the rear lens assembly but the ‘core’ section that holds the elements appears to have glue used as part of the locking ring assembly process.That must wait until tomorrow for further investigation.
From what I can see this is just the standard rear two lens elements of a three element lens assembly. The 2 lens element module that sits in the cameras mount has its own factory set focus that is locked by a very small grub screw. The removable lens barrel part of the optical system contains a single lens. The distance between the front objective element and the cameras lens mount is set using the manual focus adjustment helicoid within the lens barrel.
So in précis, if you buy an AGEMA THV4xx, 8xx or 9xx series lens. You are buying just the objective part of the lens system and are missing the two rear lens elements. Also bear in mind that these complete lens assemblies were designed to illuminate a tilting mirror that formed the first stage of the cameras raster scan system. They were not illuminating a staring array.
Fraser
-
Thanks Fraser - that is incredibly useful and interesting stuff.
Now all I have to do is work out whether the lens from that Thermovision 900 will actually do the trick and let me see through the big lens!
Even when the country is in lockdown I don't have enough time...
-
@Ultrapurple,
That THV900 lens mount likely contains a lens ‘module’ like one I found in the THV470 mount. I would expect the THV900 module to work with the THV4xx and 8xx lenses. AGEMA seem to have used the same lenses across models, just with different mounts. That makes sound business sense for a manufacturer and Inframetrics did similarly with their lenses.
The two lenses in the module are likely to be both positive meniscus types and the complete lens assembly is basically a keplerian inverting telescope with Collimating lens element at the output.
Fraser
-
If I can open the lens ‘module’ I will measure the FL of the lenses.
Fraser
-
Efforts to open the AGEMA lens mount module have failed. I used solvents and as much force on the lens ring removal tool as I dare. Greater force will likely either end up with damage to the module or me donating skin and blood to the task :scared:
The lens ring is locked in place with a caulking compound that resists all common solvents, including Acetone. I regret that is as far as I can go with the module.
I attach pictures of the lens mount and the module that fits inside it. The lens module is fine threaded and this is the focus adjustment that is set at the factory. The user focus is on the removable objective lens barrel.
As can clearly be seen from these pictures, the lens elements have succumbed to moisture that has caused corrosion of the Germanium and lifts the AR coating. This is what causes the flaking that can be seen on some used lenses. It is not repairable without expensive rework so this lens set would normally considered scrap. Interestingly, it still produces a good image !
That just goes to show how tolerant QVGA thermal cameras can be. On small spans I suspect the corrosion would be detectable.
Fraser
-
Thanks for trying - and for the excellent photos.
-
As can clearly be seen from these pictures, the lens elements have succumbed to moisture that has caused corrosion of the Germanium and lifts the AR coating. This is what causes the flaking that can be seen on some used lenses. It is not repairable without expensive rework so this lens set would normally considered scrap. Interestingly, it still produces a good image !
That just goes to show how tolerant QVGA thermal cameras can be. On small spans I suspect the corrosion would be detectable.
You might find that this is due to the lens surface not being in focus - any degradation is effectively spread over a wide area, so you may see an overall loss of transmission but not any specific bad areas. I'm not surprised it still gives a good image!
I refer everyone to this classic blog post for an example (Roger's posts are basically all worth reading btw):
https://www.lensrentals.com/blog/2008/10/front-element-scratches/ (https://www.lensrentals.com/blog/2008/10/front-element-scratches/)
-
@Hydron,
Very interesting. I am one who normally worries about dirt or minor scratches on a lens. I will be a little more relaxed about it now :-+
Fraser
-
Thanks to Hydron for finding that post. I'd seen it (or something very similar) before and was amazed at the way the lens still managed to function without problems.
I have a second-tier-brand macro lens for my DSLRs that I bought cheaply because it has fungus (https://petapixel.com/2016/09/29/remove-fungus-lens/)behind the front element - and yet it still gives images essentially indistinguishable from another, very expensive on-brand special purpose macro lens I have that's of the same focal length. (The latter is an UV-Nikkor 105mm f/4.5 - good luck if you can find one: few were made, they went out of production last century, and it took me six years to find one that cost less than five grand).
From the above-linked Petapixel page:
(https://petapixel.com/assets/uploads/2016/09/img_3064-edit-small.jpg) (https://petapixel.com/2016/09/29/remove-fungus-lens/)
The fungus is the fine, thread-like whiteness in the centre of the lens. Mine was a different type of lens, and the fungus was on the 'other end'.
The bottom line is that I don't worry too much about thermal lenses that have defects, particularly in the coatings. True, a perfect lens will work better than one where the coating is flaking off by the handful, but we work at such low resolutions (and with so many other compromises in lenses trying to work over a 2:1 wavelength range) that the difference is marginal in most instances. No, I wouldn't be happy using a flaked lens for exacting science work, but for everything else they're fine.
-
Ultrapurple,
You are so right. We are generally not using our cameras for exacting science and it is all too easy to become fixated on perfection when such is not justified. I am a perfectionist with a bit of pessimist thrown in for good measure so strive for perfection whenever possible. Sadly when I see a defect in something I always want to make it better or replace it. Where lenses are concerned, we are often indoctrinated with the careful handling of such to avoid damage or delicate surfaces. Every SLR lens I ever bought had a UV filter fitted to it at the time of purchase..... it seemed the sensible thing to do at the time. Pro photographers have highlighted the downside of such filters though and there is a definite split in the photography community on the topic of lens protection filters.
I have previously posted comments and articles from FLIR regarding the issues of dirt on external thermal CCTV camera windows and lenses. They say that dust and minor lens contamination is not as degrading to a thermal image as some might think. They recommend against wash wipe systems as they do more harm than good.
Sadly when looking at a lens that operates in another area of the EM spectrum we only see the VL spectrum properties of the surface. These properties can be very different to those seen by the sensor that is operating in the intended band. A lens can look darned ugly in VL yet operate fine in MWIR or LWIR.
One thing I do know for sure, if a lens has a damaged AR coating, it is better to leave well alone rather than attempt total removal of the AR coating to make it look ‘pretty’ !
I noted evidence on some lens elements being sold on eBay of a seller having polished away the AR coating of MWIR lenses to make them work over a broad spectrum? That is not something I would recommend. Transmission is less than 46% but the lenses were left heavily scuffed by the process.
Fraser
-
I gave up on using protective (sacrificial) filters on my DSLR lenses when I discovered just how badly they affected the images. Although (as we've seen in the above reference) a smashed front element can still produce a good image, I found that putting anything in front of my lenses robbed them of sharpness. Initially I thought it was something like the autofocus on that specific lens or body getting thrown off, but I found the same thing happened with different lenses, filters, and bodies. I like my subjects tack-sharp and I just don't get that with a filter. Of course, sometimes I have to swallow my pride a bit (for example with graduated filters, or high-opacity neutral density filters for long exposures) and of course my near-IR and UV photography can only be done with filters. But, by and large, I am prepared to take the (small) risk of damage to the front element of a lens in exchange for images that please me better.
(https://live.staticflickr.com/960/41035074805_ee86799f87_o_d.jpg) (https://www.flickr.com/photos/ultrapurple/41035074805/)
Small Spanish lizard, taken from about 2m away, Nikon D850 and 300mm at f/8, ISO 64, 1/250s; the creature was only about 100mm long so this is a crop from a much larger image. Any softness from a filter on the lens would have rendered this image unusable.
(https://live.staticflickr.com/1808/43470757401_eb281c0aba_o_d.jpg) (https://www.flickr.com/photos/ultrapurple/43470757401/)
Similar, but much closer to home; close crop from 400mm lens; cat was about 15m away.
-
Yeah I don't normally use "protective" filters either, or if I do it's only because I'm in a situation where I'm worried about getting crap on the surface. The UV/protective filters I do have were also all bought second hand for essentially pocket change from a good independent camera shop (sadly I'm now on the other side of the world from it) - they also tend to be rather pricey bought new, especially given they normally just make images worse!
As for the macro - it's probably actually hard to find a bad ~100mm macro lens regardless of brand; a F2.8 lens in the 90-105mm range isn't hard to do well and doesn't need any exotic optical materials, techniques etc. A few years I got a copy of the original 1990s 100mm Canon EF macro in absolutely mint condition (complete with original box etc) for under half the price the newer USM internally-focusing version sells for second hand and it's serving me very well. Won't do all the funky UV/NIR stuff your UV-Nikkor will though (I am aware of it's reputation - next up I guess is the coastal optics 60mm uv-vis-ir lens?).
-
@Hydron - Coastal Optics make a direct (authorised) clone of the UV-Nikkor 105mm f/4.5. Last time I looked it was about US$8000 plus shipping and local taxes - so about £8000.
-
Yeah that rings a bell - I know the UV-Nikkor was still available somehow (I think it was used industrially too), but their 60mm apochromatic lens is a fresh design and is apparently basically unmatched in performance (probably in price too unfortunately).
One day I'll end up modding an old camera to remove the IR cut filter and have a play with near-IR photography, but certainly not interested enough in UV/near-IR to spend UV-Nikkor kinda money.
-
Nikon made a 55mm f/4 in the mid-1960s. Details of it here. (https://www.mir.com.my/rb/photography/companies/nikon/nikkoresources/special/55mmUV.htm) That page also says Nikon also launched a 55mm f/4 UVIR lens in 1988. There were other low-run specials (eg for NASA, a 55mm f/2 UVIR for space use). The page I linked above has pretty much all that's known; the wider Mir.com site has a great deal of authoritative info on virtually every Nikon lens ever produced (including rare things like the 10mm f/5.6 OP FISHEYE-NIKKOR: Grays of Westminster currently has one on offer for GBP30,000 - and that's a lens from about 1970).